# Gemini 2.5 Flash Lite

The entry point to Gemini 2.5's benchmark improvements, configurable thinking at the lowest cost in the 2.5 family, with first-token latency lower than any other 2.5 model.

- **Model ID:** `google/gemini-2.5-flash-lite`
- **Type:** chat
- **Providers:** vertex, google
- **Context window:** 1,048,576
- **Maximum output tokens:** 65,535
- **Pricing:** $0.1/1M input tokens, $0.4/1M output tokens
- **Canonical page:** https://vercel.com/ai-gateway/models/gemini-2.5-flash-lite

## Supported parameters

Detailed capability metadata has not been reported for this model.

## Example

```ts
import { streamText } from 'ai'

const result = streamText({
  model: 'google/gemini-2.5-flash-lite',
  prompt: 'Why is the sky blue?'
})
```

## About

Gemini 2.5 Flash Lite is the efficiency tier of the Gemini 2.5 family, released June 17, 2025 alongside 2.5 Flash and 2.5 Pro going to general availability. It runs faster and costs less than any other 2.5 model while outperforming 2.0 Flash-Lite on benchmarks that matter for real-world developer tasks: coding, mathematics, scientific reasoning, and instruction following.

Configurable thinking is the feature that most distinguishes Gemini 2.5 Flash Lite from 2.0 Flash-Lite. At inference time, you set a thinking level (minimal, low, medium, or high) to allocate more deliberation to harder problems without switching endpoints. This is the same thinking mechanism available in 2.5 Flash and 2.5 Pro, scaled down to the lite budget. For tasks that occasionally need more reasoning depth than a pure speed-first model provides, the thinking toggle avoids the cost jump of routing to a full reasoning model.

For teams running 2.0 Flash-Lite in production and evaluating a 2.5 upgrade path, Gemini 2.5 Flash Lite is the migration-friendly choice: better benchmark performance, thinking capability, and latency that matches or beats the previous generation.

## What to consider

Applications using the thinking feature should benchmark total token cost under realistic thinking budgets, as thinking tokens contribute to output costs.

## When to use

### Best For

**High-volume agentic pipelines needing occasional reasoning:** The thinking toggle allows selective deliberation on harder steps without paying full 2.5 Flash prices for every call in the pipeline.

**Migrating from 2.0 Flash-Lite:** Benchmark improvements across coding and math mean the upgrade delivers measurable quality gains on common developer tasks at comparable cost.

**Latency-sensitive applications within the 2.5 family:** When 2.5 Flash or 2.5 Pro latency is too high for the user experience, Flash-Lite provides 2.5-generation quality at the fastest 2.5 response times.

**Translation, classification, and data extraction at scale:** Strong instruction following and fast response make it a reliable workhorse for structured-output production tasks.

### Consider Alternatives When

**Maximum reasoning depth is required:** 2.5 Flash or 2.5 Pro with uncapped thinking budget is more appropriate for the most complex multi-step problems.

**Image generation is needed:** Gemini 2.5 Flash Lite does not generate images. Gemini models with native image output are available in the 2.5 Flash Image and 3.x families.

**Your workload is pure annotation/extraction without reasoning:** For text-output-only extraction at maximum cost efficiency, 2.0 Flash-Lite's lower price floor may be preferable.

## Best for

- **High-volume agentic pipelines needing occasional reasoning:** The thinking toggle allows selective deliberation on harder steps without paying full 2.5 Flash prices for every call in the pipeline
- **Migrating from 2.0 Flash-Lite:** Benchmark improvements across coding and math mean the upgrade delivers measurable quality gains on common developer tasks at comparable cost
- **Latency-sensitive applications within the 2.5 family:** When 2.5 Flash or 2.5 Pro latency is too high for the user experience, Flash-Lite provides 2.5-generation quality at the fastest 2.5 response times
- **Translation, classification, and data extraction at scale:** Strong instruction following and fast response make it a reliable workhorse for structured-output production tasks

## Consider alternatives

- **Maximum reasoning depth is required:** 2.5 Flash or 2.5 Pro with uncapped thinking budget is more appropriate for the most complex multi-step problems
- **Image generation is needed:** Gemini 2.5 Flash Lite does not generate images. Gemini models with native image output are available in the 2.5 Flash Image and 3.x families
- **Your workload is pure annotation/extraction without reasoning:** For text-output-only extraction at maximum cost efficiency, 2.0 Flash-Lite's lower price floor may be preferable

## Frequently asked questions

### What thinking levels does Gemini 2.5 Flash Lite support?

Four levels: minimal, low, medium, and high. You set the level per request. Thinking tokens are added to output token count, so higher thinking levels increase both quality and cost.

### How does Gemini 2.5 Flash Lite compare to 2.0 Flash-Lite in benchmark performance?

Gemini 2.5 Flash Lite shows cross-category benchmark improvements in coding, mathematics, science, and reasoning over 2.0 Flash-Lite.

### Does Gemini 2.5 Flash Lite support image and audio inputs?

Yes, the model accepts multimodal inputs including images, audio, and documents alongside text, within the context window of 1.0M tokens.

### What is the latency profile compared to 2.5 Flash?

Gemini 2.5 Flash Lite is the fastest and lowest-latency model in the 2.5 family. It provides 2.5-generation capability with first-token times lower than 2.5 Flash or 2.5 Pro.

### When does it make sense to use thinking in Gemini 2.5 Flash Lite?

When a subset of requests in a pipeline hits problems that need more deliberation, such as math problems, multi-step instructions, or ambiguous classification. Setting thinking to minimal for routine requests and medium for flagged hard ones keeps average cost low.

### How do I use Gemini 2.5 Flash Lite on AI Gateway?

Use the identifier `google/gemini-2.5-flash-lite` with any supported interface. Set thinking level via provider options in the AI SDK or via request parameters in direct API calls.

## Links

- [Model page](https://vercel.com/ai-gateway/models/gemini-2.5-flash-lite)
- [AI Gateway documentation](https://vercel.com/docs/ai-gateway)
- [Provider model documentation](https://cloud.google.com/vertex-ai/generative-ai/docs/models/gemini/2-5-flash-lite)
- [Provider pricing](https://cloud.google.com/vertex-ai/generative-ai/pricing)